Automated Summarization of Lecture Videos
Anusha Vimalaksha, Siddarth Vinay, Abhijit Prekash, N. S. Kumar · 2018
Currently, educational institutes record videos of lectures for online courses, as well to allow students to watch the videos at a later time. Most of the videos are for a fixed duration with very little indication of the subtopics. These videos also consume large amounts of backup memory space over a period of time. Our innovative tool provides a mechanism to break the video into multiple parts based on crowd sourcing. The participants note the topic and the time stamps of beginning and end of a topic in an interactive session. Our tool breaks the video based on the time duration and creates new videos titled based on the subtopic. This tool takes a few seconds – approximately 17 sec – to process one hour video – so that all videos recorded in a single day may be processed in an hour. Our tool also provides a mechanism to summarize the video so that the important aspects of the video are captured for viewing as well for archiving. We get the audio transcribed and do a summarization of the resultant text and match this back to the video. This reduces the length of the video by approximately 50% and saves storage space by the same amount. This has been tested with videos recorded in our University.